Evaluating Behavioral Interventions Week 3:Interpreting & Graphing Data
Benefits of Constructing Graphs Immediate updates Variability Practical interpretation Conservative Method Encourages Personal Judgments Feedback to client
Components of Graphs X axis (time – and/or independent variables) Y Axis (Behavior- Dependent Variable) Phase lines (ABC’s) Levels (1,2,3’s) Labels (phase and conditions) Data Points Data Path (connection of data) – Lines – Arrows – Breaks Figure Caption
Types of Display Frequency Rate Cumulative Record Latency Duration
Semi-logarithmic graphing Rate 2-4 is doubling increase Rate of is doubling increase Solution? – Put Y axis on equal intervals
Celeration Charting Semilogrithmic on Y axis AND six X10 cycles – * means acceleration – X means deceleration
Visual Analysis Number of data points Variability Level – Mean – Median – Percentage of Non-overlapping Data points (PND) Trend – Free Hand (not recommended) – LSLR (recommended – need some long hand calculations) – Split-middle line of progress (split data in half, find mid rate and mid date in both panes and connect) Quarter Intersect probably close enough!
A word about statistics Means are often inappropriate Stat analyses typically violate assumptions – E.g. autocorrelations Variability is an issue Flexibility in the design is needed in applied environments Component Analyses are often needed
Errors: A word about Visual Analysis If a problem is solved you should see it. You don’t need to rely on stats! Visual analysis good for finding Robust, reliable large changes Low Type I Error (False Positive) Higher Type II Error (False Negative)
Designs A-B A-B-A A-B-A-B A-B-A-B-C A-B-A-C-B-B2 Multiple Baseline Designs Alternating Treatments Design Changing Criterion Designs